Artificial Intelligence has been threatening to transform the contact centre. But how much of this talk is hype? And how much is likely to happen?
Could this lead to the decline of contact centres as we know them?
During this webinar we debunked the myths from the reality.
- Introductions – Jonty Pearce, Call Centre Helper
- Martin Hill-Wilson- Brainfood Extra, Ashley Unitt– NewVoiceMedia and Tim Pickard – NewVoiceMedia
- Artificial Intelligence
- Examples where Artificial Intelligence is working
- Robots and Chat Bots
- Speech Recognition
- Neural Networks
- Where does software end and Artificial Intelligence begin?
- What will be the impact on jobs?
- Will customer service go down?
- Search engines vs. Chat Bots
- How fast will it happen?
- Top tips and projections from the audience
- Winning Tip – “So what you can do with AI is based on the data it’s fed. Therefore, making sure AI can access the right data is paramount to those looking to implement AI soon. Also, if customer conversation and transaction data is part of that, getting historical data is beneficial as well (more data = better problem solving).” Thanks to Thomas13.
Original Webinar during our Online Conference: Tuesday 28th March 2017
Questions and Answers
1. Question relating to market readiness: Wouldn’t this depend on your market demographic? i.e. if they aren’t the early adopters you probably have a few years before they want/expect A.I.
There are many forms of AI. It may take some time for your customers to accept automated responses or virtual assistants, however this does not mean you cannot be using AI in the background to assist and augment your agents and be getter better insights on how your business operates or what your customers are saying. We believe there will be rapid adoption of AI in businesses as part of their processes, even if it’s largely invisible to customers. Adopting AI will give large advantages to companies and hence will be driven by completion, you’ll need AI to be able to compete.
2. So in simple terms… is AI just a nice use of the knowledgebase? (it’s output)
Enhanced use of knowledge bases is part of AI but AI is so much more. There are many forms of AI that learn from experience and from many other forms of data. Where there is availability of good data then AI can be used to help make many decisions and at times this can be better than human decision making. Where there are reasonably well defined problems or limited fields of knowledge then AI can automatically perform many tasks. The scope of these tasks is widening at a fast rate.
3. What did you say the ex-Google employee who said that speech recognition is at 99% was?
Andrew Ng. He has predicted that speech recognition will soon get to 99% accuracy. Currently Andrew’s Deep Speech 2, developed at Baidu, is at >= 96% accuracy which is better than most humans.
4. How is speech readiness available across other than english?
Most research tends to be focussed on English and Mandarin due to market size and easier availability of training data. Companies like Google, who have the advantage of access to large amounts of speech data, have wide language support but the availability from other vendors is limited. Google supports 80 languages well.
5. To what extend AI is already used to pre-empt issues and improve the primary process avoiding rather than helping to fix issues?
Many companies are already using AI to analyse customer sentiment and queries by many large companies to help avoid issues or diagnose them quickly. However these tends to be very bespoke solutions requiring large investment and manual configuration. The next wave of AI will democratise this access and bring this functionality to many companies in easier to use solutions.
6. What is the outlook if we are using AI to contact customers B2B and the customers are using AI to deal with this kind of lead generation for example? Are we humans going to loos touch of what is going on?
Currently these virtual assistant to virtual assistant communications are some way off. Probably the first examples of this will be assistants communicating to be able to find mutually convenient times to meet etc. It will take time for humans to build confidence in VAs to be able to feel comfortable to be able to delegate more and more of their lives to VAs. This will happen over time and there will be many mistakes made along the way and many scare stories but it will eventually happen. Delegation will inevitably lead to some loss of control but will be ultimately empowering, however there will need to be many discussions about the ethics of giving more control of your lives to VAs ultimately created by third parties such as Google. In many areas the technology of AI is moving more quickly than our laws and ethics and this will lead to mishaps and some inevitable backlash before we all work this out.
7. From a business point of view there are obvious advantages of AI within a modern contact centre but has there been much market research with the customers on their preference for AI or human contacts?
Last year NewVoiceMedia polled 1,000 US citizens about their preference for AI or human contact, and uncovered that customers are not ready for a complete AI service overhaul.
68 percent of respondents claimed they would prefer to interact with a live agent rather than automated self-help (FAQs/guided support, dial directories, chatbots, etc.) when dealing with customer service, and 63 percent felt that personalisation should be the focus of improvement in customer service.
It’s not that technological advancements cannot improve the customer experience. They absolutely can. But companies must first have a foundation of good support built on fundamentals that include phone functionality and in-depth customer insights. Find more details about the research at